If you analyze gas price differentials like most everyone else, then you may be underestimating the benefit of increasing gas prices.

Most people (including me historically) use a single average in a single cell in EXCEL to determine the netback price of natural gas. Due to the time-intensive data handling requirements for analysis of lease operating statements, we often analyze price differentials using only one year of data and strictly as a ratio to benchmark prices; the average is the answer with perhaps a review of the array of values.

Granted, we often can't get (or don't have time to read) the full gas sales contracts which spell out explicitly the parameters of the sale when we are making an evaluation. We do know that those contracts can outline several layers of provisions, but they commonly include at least these characteristics:

- specification of a reference price,
- a component of cost variable with the reference price,
- a fixed cost component, and
- provisions for step-changes in the terms of the contract.

The methods we usually use simplify the analysis to only value--the variable component--and it lumps or ignores all of the other parameters. Looking at a column of ratios (or worse, a row) obscures and oversimplifies the patterns. If we, by analogy, tried to forecast reserves with a regression applied formulaically to a series of twelve numbers in a line, we would be summarily dismissed. The table below demonstrates how we commonly do just that with our conventional analysis of short term pricing data.

A simple graph of ratio versus time would expose at least one additional variable, namely step changes, but the better answer is a two-dimensional graph and including several years of data.

By cross-plotting netback price against benchmark price, the analyst can identify with better accuracy both the variable and the fixed costs of the gas contract, and by coloring the data points by date, we can expose step changes in the contract terms.

The common practice of averaging the ratios for each month is equivalent to fitting a line with the intercept forced to zero as shown in the graph. When prices are low, the fixed cost component brings down disproportionately the implied ratio to benchmark. Using this fit to determine netback prices at higher benchmarks undercalls the price and thus understates the value. The conventional method cuts the other way when prices go down, it predicts higher prices than would be received when prices fall.

This kind of analysis, like most any, benefits from more data. As long as the data covers a reasonable range of realized prices, it can be regressed with a line whose slope is the variable component and intercept is the fixed cost component. Prices in the last twelve months do cover such a range, but data for a more stable period might not show enough variation in benchmark prices to refine the fit as uniquely.

Though superior, this kind of analysis has mostly not been deployed for reserves and economics calculations due to the difficult of getting and handling data. Truly, the best answer is for programs like PHDwin and Value Navigator to build LOS analysis into their engines. They already have database engines, graphing engines and calculation engines, but I have not yet persuaded them to build the features. Truly, our reserves & economics programs should use all the technical features for the economics as much as for the reserves. (Please tell them I said so!)

In the meantime, Philip Anthony and I have almost finished deploying an analytical tool much easier and more thorough than the standard, from-scratch analysis of each spreadsheet of LOS data. You can watch LinkedIn or follow the newsletter to learn more about our Mass LOS solution.